cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2061
Views
3
Helpful
8
Replies

"The Message Server is offline" when on Exchange 2007 CCR 2nd node.

Ryan17
Level 1
Level 1
After rebooting our Unity Server I still had the same behavior.
I then, on a desperate hunch, failed over our Exchange 2007 cluster CCR which had been running on the "2nd" node due to maintenance performed last night. Once I failed oner to what is normaly the active node everything started working.
Users can now enter their PW on the phone and get their Vmail in their inbox.
I confirmed that when I fail back over to the other node and attempt to check my voicemail I get a "The Message Server is offline" message.
All configuration in the "permissions Wizard" points to the cluster name and not the node.
Any Idea what might be wrong.
<<We re-paired Unity with the 2007 server about a month ago.)
1 Accepted Solution

Accepted Solutions

Yes, that is correct. If a mixed environment, Unity has to be partnered with the Exchange 2007 server. Honestly, I would schedule some downtime and do a full resync of the directory via the message store and test functionality when failing over to the backup Exchange server. Here is an excerpt of how that should occur:

OK – so the next step is, IMO or from what I’ve been able to find, not well documented. The first time I had to do it, I had that “what the heck is this going to do” feeling. However, it’s not so bad. Here’s where you manually force Unity to sync using the Message Store Configuration wizard. To do this, you must be logged in as a user with rights to the mailstore – personally, I use the UnityInstall account for this task.

7. Open a command line prompt and navigate to the Commserver folder on your server. In my example, I’ve loaded Unity on the D: drive so we’ll be working from that root. So, type D: and enter.

8. Type cd commserverconfigurationsetup and enter.

9. Now type setup.exe sync.

10. The Cisco Unity Message Store Configuration Wizard launches full-screen. Click Next.

11. A process runs and provides a status icon. In short, it says to please wait while Cisco Unity fully resynchronizes with the mailstore. Do not click Next. This process could take quite a long time depending on the size of the database and other factors. For reference, I did this for a customer with about 4000 subscribers and it took approximately 5 hours to complete.

12. When the sync is complete, the Wizard proceeds without prompting and indicates that it is restarting services. When done, click Finish.

If you’re interested to know what goes on in the background, you can take a look at the application log. You’ll notice that the primary directory replication and SQL change writer services stop and start. Once complete, some folks prefer to reboot the Unity server – others do not. Personally, I don't (or have not yet). However, I have seen that it can take a while for things to completely clear up. My recommendation is to create a test subscriber, wait 15-20 minutes for AD, and then verify if the SMTP address is stamped in AD and then in SQL. If it’s not written to SQL right away, be patient and check again within a few hours…you may find it just took a while. Of course, if you still have issues after performing these procedures…it may be time for a TAC case.

The original blog post is related to another issue, but is here for you to browse if you wish. http://www.netcraftsmen.net/resources/blogs/resynchronization-of-unity-with-the-directory-using-the-message-store-configuration-wizard.html?blogger=David+Hailey

Hailey

Please rate helpful posts!

View solution in original post

8 Replies 8

David Hailey
VIP Alumni
VIP Alumni

This is a duplicate post.  I responded to your other post in the Video over IP forum.  Most of us check all forums, so need to double up postings.

Ryan17
Level 1
Level 1
After rebooting our Unity Server I still had the same behavior.
I then, on a desperate hunch, failed over our Exchange 2007 cluster CCR which had been running on the "2nd" node due to maintenance performed last night. Once I failed oner to what is normaly the active node everything started working.
Users can now enter their PW on the phone and get their Vmail in their inbox.
I confirmed that when I fail back over to the other node and attempt to check my voicemail I get a "The Message Server is offline" message.
All configuration in the "permissions Wizard" points to the cluster name and not the node.
Any Idea what might be wrong.
We re-paired Unity with the 2007 server about a month ago.)

David Hailey
VIP Alumni
VIP Alumni

In relation to the Unity server, where is the primary partner server? How about the Exchange backup node? Unity is AD site aware so do you have AD Sites and Services properly configured as well? Is there possibly a firewall between the Unity server and the backup Exchange box?

Hailey

Please rate helpful posts!

I tried to re-edit my post and now I just have the title showing. I hope you can read this.

The Exchange 2007 cluster and Unity are in the same site.

I found the problem is fixed when I move the Exchange Cluster Resource to the node that was Active when I "paired" it with Unity.

Move it back and I get "The message Server is offline" So, thankfully Voicemail is working for everyone now and I have a better idea where the problem is.

Now when running the permissions wizard everything related to exchange was named by its cluser name MAIL as opposed to an individual cluster node name, Mail-Node1, Mail-Node2.

Where did I go wrong?

-Ryan

I follow the issue. You said you repaired the Unity server to Exchange a couple months ago...can I ask why and how you did that? I have a customer running UM with Exchange 2007 CCR cluster and have no known issues there. I have a procedure to resync Unity with the mailstore; however, I'm not sure you need to do that again. What version of Unity are you running?

The original Exchange 2003 cluster that Unity was paired withis located in a Colo DataCenter we are shutting down.

The new Exchange 2007 server is in our local DataCenter and now local to the Unity Server.

Once I moved a few test users to the 2007 server they could not get there voicemail.

Working with Cisco TAC they instructed me to -re-pair with the 2007 server and gave me instructions on running the Cisco Message Store Configuration Wizard. Everything seemed to work fine with this and after the change users on both the 2003 and 2007 servers recieved voicemails fine.

Now, this issue has just arrisin as it is the first time I've probably been running exchange on this node during the day.

again, the rest of exchange seems happy.

Thank you for your help!

-Ryan

Yes, that is correct. If a mixed environment, Unity has to be partnered with the Exchange 2007 server. Honestly, I would schedule some downtime and do a full resync of the directory via the message store and test functionality when failing over to the backup Exchange server. Here is an excerpt of how that should occur:

OK – so the next step is, IMO or from what I’ve been able to find, not well documented. The first time I had to do it, I had that “what the heck is this going to do” feeling. However, it’s not so bad. Here’s where you manually force Unity to sync using the Message Store Configuration wizard. To do this, you must be logged in as a user with rights to the mailstore – personally, I use the UnityInstall account for this task.

7. Open a command line prompt and navigate to the Commserver folder on your server. In my example, I’ve loaded Unity on the D: drive so we’ll be working from that root. So, type D: and enter.

8. Type cd commserverconfigurationsetup and enter.

9. Now type setup.exe sync.

10. The Cisco Unity Message Store Configuration Wizard launches full-screen. Click Next.

11. A process runs and provides a status icon. In short, it says to please wait while Cisco Unity fully resynchronizes with the mailstore. Do not click Next. This process could take quite a long time depending on the size of the database and other factors. For reference, I did this for a customer with about 4000 subscribers and it took approximately 5 hours to complete.

12. When the sync is complete, the Wizard proceeds without prompting and indicates that it is restarting services. When done, click Finish.

If you’re interested to know what goes on in the background, you can take a look at the application log. You’ll notice that the primary directory replication and SQL change writer services stop and start. Once complete, some folks prefer to reboot the Unity server – others do not. Personally, I don't (or have not yet). However, I have seen that it can take a while for things to completely clear up. My recommendation is to create a test subscriber, wait 15-20 minutes for AD, and then verify if the SMTP address is stamped in AD and then in SQL. If it’s not written to SQL right away, be patient and check again within a few hours…you may find it just took a while. Of course, if you still have issues after performing these procedures…it may be time for a TAC case.

The original blog post is related to another issue, but is here for you to browse if you wish. http://www.netcraftsmen.net/resources/blogs/resynchronization-of-unity-with-the-directory-using-the-message-store-configuration-wizard.html?blogger=David+Hailey

Hailey

Please rate helpful posts!

Wow! Thanks for the answers!

I will try this this evening.

Part of me is wondering how this is possible, given the Market share Cisco and Microsoft has.

I guess one main reason I try and deploy everything in a redundant manor is so I don't have to sweat, or rush do something manualy, when a failure occurs.

(especialy when it is a gracefull failover)

I guess I am surprised as I have never had similar issues with other systems. Shoretel, 3com.

Given the difference in price point and market exposure I'd expect to have these kinds of high availability issues addressed.

Once again you've been a huge help. thank you.

-Ryan

I'll post my results in the morning.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: